News
Designing Molecular Photoswitches with Generative Machine Learning
Robert's work on the design of molecular photoswitches with generative machine learning (ML) just came out in Digital Discovery. Designing photoswitches (molecules that respond to light by changing their structure) is highly complex, because different properties must be optimized at the same time. For example, depending on the application light must be efficiently absorbed in a certain regin of the spectrum, and the switched state should fulfil some stability criteria. The multi-objective generative AI model that Robert designed allows balancing such criteria in a Pareto optimal way. Together with the group of Stefan Hecht at HU Berlin, some of the predicted molecules were then synthesized and characterized, revealing new design rules for Spiropyran based photoswitches.